What the first full week across millions in spend tells us about budget limits, conversion quality, and what advertisers should watch next.
Google’s August 17 change was designed to make campaigns using Target CPA or Target ROAS, particularly those limited by budget, perform more consistently toward the targets advertisers set. In Google’s example, a campaign with a $10 Target CPA that had been delivering at a $5 actual CPA could begin moving closer to $10 unless the target changed.
Well ahead of August 17, our teams reviewed the campaigns most likely to be affected, compared Target CPA and Target ROAS settings with recent actual performance, and checked which campaigns were limited by budget. Where changes were needed, we made them in measured steps or through controlled tests. In other cases, the right decision was to leave a stable campaign alone.
Now we have the first full week of data. It is too early for a final verdict, but it is enough to see a few clear patterns. This is not an Amsive-only event. Every advertiser entered the same auction with different targets, budgets, and levels of preparation. As brands and agencies react, the auction is still settling.
Amsive analyzed millions of dollars in Google Ads spend across dozens of active advertisers and hundreds of campaigns spanning financial services, healthcare, home services, eCommerce, education, B2B, insurance, senior living, hospitality, and consumer services. We compared August 17–23 with the average of the three prior full weeks, July 27–August 16, so one unusual week did not become the entire story.
What the First Week Adds Up To
- Conversion quality, not CPC, was the bigger Search issue. Search CPC increased 4%, while conversion rate fell 15% and CPA increased 21%.
- Budget status changed how the shift showed up. Budget-limited campaigns paid more for fewer clicks, while campaigns with room to spend found cheaper traffic that converted less often.
- Target CPA and Target ROAS did not react the same way. Target CPA Search stayed relatively close to baseline; Target ROAS got more clicks at a lower CPC, but ROAS declined 32%.
By vertical, the headline was different:
- Financial services: Higher CPA was the clearest issue.
- Home services: Both CPC and CPA increased, with wide differences by DMA.
- Healthcare: Traffic became cheaper, but conversion rate fell.
- eCommerce + retail: ROAS was the clearest concern in the value-based campaigns we reviewed.
- B2B technology: CPA increased, making lead quality and pipeline conversion even more important.
- Education: Performance was less volatile and stayed close to the pre-change baseline.
- Insurance: It was one of the more stable categories in the first week.
- Senior living: CPA increased modestly, while tour quality, move-ins, and community capacity remained the more important business measures.
The Bigger First-Week Story Was Conversion Quality, Not CPC
Across the comparable Search campaigns, spend and click volume were nearly flat. The bigger change happened after the click.
The gap between CPC and CPA is the point. Click costs changed only slightly, but Search traffic converted less often, pushing CPA higher. If CPC and lost impression share to rank are rising, review Target CPA or Target ROAS settings, budgets, and auction competitiveness. If CPC is stable or falling while conversion rate drops, look at search terms, query mix, landing pages, and the conversion signals guiding Google.
Platform CPA also cannot be the last word. A lead form, an appointment, a purchase, and a qualified sales opportunity are not interchangeable. Google can report a conversion while the business receives a very different result.
Brand Clicks Cost Less, but Converted Less Often
Separating Search campaigns by Brand, Generic, and Competitor intent reinforced the same point: higher CPA was not always being driven by higher CPC.
Brand clicks were slightly cheaper, but the lower conversion rate pushed CPA higher. Before changing bids, check whether the brand query mix, conversion tracking, landing pages, or user behavior changed.
Generic Search followed a similar pattern, though the change was less pronounced. That is where search terms, query expansion, negative keywords, and landing-page alignment deserve a closer review. Competitor campaigns improved in the first week, but the volume was lower, so we are monitoring the trend rather than treating it as a broad conclusion.
Exact Got More Expensive. Broad and Phrase Converted Less Often.
Breaking Search out by Broad, Phrase, and Exact match showed two different issues: Exact became more expensive, while Broad and Phrase converted less often.
Broad Match still deserves close attention because Google has more freedom over both what it pays and which searches it enters. But the answer is not to turn Broad Match off. Modern Search management is increasingly an exercise of exclusion rather than inclusion. Google is designed to go broad. The advertiser’s job is to shape what the system learns through strong conversion signals, negative keywords, clear campaign structure, and testing that matches the business’s risk tolerance.
Related Amsive reading: The AI-Powered Paid Media Playbook with Google and 3 AI-Optimized Frameworks to Improve Your Google Campaigns.
Budget-Limited Campaigns Paid More. Campaigns With Room to Spend Converted Less Often.
Google designed the August 17 change around target-based campaigns that were limited by budget. That makes budget status central to this analysis, not just another way to segment the data. In the broader Search group we reviewed, nearly four out of every five post-change dollars were spent in campaigns Google labeled Budget constrained.
Budget-constrained campaigns paid more per click and received fewer clicks. Their CPA increased, and the shift from lost impression share to budget toward lost impression share to rank points to greater auction competitiveness within the money available.
Campaigns with room to spend found more clicks at a lower CPC, but conversion rate fell sharply and CPA still increased. Google found more traffic at a lower price, but that traffic converted less often.
Google’s Budget constrained label can change from week to week, so this is a directional read rather than proof of cause. In future installments, we will keep the same pre-change groups in place and watch whether the pattern holds.
Target CPA Stayed Close to Baseline. Target ROAS Got More Clicks but Lower Return.
Search campaigns using Target CPA stayed relatively close to the pre-change baseline in the first week.
Target ROAS delivered more clicks at a lower CPC, but those clicks produced less value and ROAS declined. The value-based group is smaller, so this is an early signal rather than a market-wide benchmark, but it is worth watching.
The takeaway is not that Target CPA is good and Target ROAS is bad. The target, the quality of the conversion value sent to Google, and the business model all matter. Google can optimize toward the value it receives. It cannot decide whether the advertiser defined that value correctly.
What Changed by Vertical + What to Check Now
Financial services: CPA was the clearest challenge
Across the comparable financial services accounts, CPA was roughly 30% higher than the pre-change baseline. High-intent banking, deposit, lending, and credit campaigns already operate in competitive auctions, so even a modest shift in bidding can add cost quickly.
What to check now: approved application rate, funded accounts, acquired customers, and revenue by product. A lower-cost lead is not useful if approval or funding rates fall.
Home services: CPC and CPA increased, with wide variation by DMA
Home services CPA increased about 22%, and CPC also rose, but the size of the change varied widely by market. The same service line can behave very differently from one DMA to another because demand, competition, capacity, and lead value are local.
What to check now: booking rate, job rate, revenue, and contribution by market and service type. A single national Target CPA policy can miss what is happening in individual markets.
Healthcare: Traffic became cheaper, but converted less often
Healthcare CPC declined about 10%, while conversion rate fell about 31% and CPA increased about 29%. The lower click cost did not translate into more efficient patient/enrollment acquisition.
What to check now: the mix of calls, forms, provider searches, appointments, and patient actions, along with appointment quality and patient acquisition outcomes. These actions do not all carry the same value.
eCommerce + retail: ROAS was the clearest concern
Across the comparable Target ROAS retail campaigns we reviewed, ROAS declined by roughly 29% to 35%. Because this is a smaller group, we are treating it as an early signal rather than a retail-wide benchmark.
What to check now: revenue per click, average order value, margin, product and category mix, and new-customer share. Those measures show whether additional spend is finding more valuable customers or simply more transactions.
B2B technology: CPA increased, making lead quality the real test
In the B2B technology accounts we reviewed, CPA increased by roughly 28% to 34%. The direction was consistent, but a platform conversion is only the start of the sales process.
What to check now: marketing-qualified lead rate, sales acceptance, opportunity creation, pipeline, and closed revenue. If Google produces more form fills than Sales does not want, platform CPA can look reasonable while business performance weakens.
Education: Performance was less volatile
Education stayed relatively steady in the first week, with aggregate CPA close to the pre-change baseline. That is an important reminder that not every vertical reacted the same way.
What to check now: applications, completed applications, enrollment, and program-level quality over a longer window. Many education campaigns have lower weekly conversion volume, so short-term changes can look larger than they are.
Insurance: Performance was comparatively stable
Insurance CPA increased about 5%, making it one of the more stable lead-generation categories in the first week.
What to check now: quote starts, completed applications, bound policies, and retained customers by product. Those should not be treated as equal outcomes.
Senior living: CPA increased modestly
Senior living CPA increased about 8% in the first week, a smaller change than several other lead-generation categories.
What to check now: tour quality, move-in rate, care type, and community capacity. A campaign can look less efficient at the lead level while still producing the right residents for locations that have room.
.
What Advertisers Should Do Now
1. Recheck the Target CPA or Target ROAS you are giving Google
If a campaign has been delivering well below its Target CPA, ask whether the target still reflects what the business is willing to pay. The same applies to Target ROAS. A target that was once treated as a loose guardrail is now a more direct instruction. Review it against recent performance and actual business value, then make changes in measured steps.
2. Review lead and customer quality on a recurring basis
For lead generation, compare qualified lead rate, appointment or booking rate, sales acceptance, close rate, revenue per lead, and customer value before and after August 17. For eCommerce, track revenue per click, average order value, margin, product mix, and new-customer contribution. These measures will tell you more than CPA or purchase count alone.
3. Identify whether the problem is cost or traffic quality
If CPC and lost impression share to rank are rising, auction competitiveness may be the main issue. If CPC is falling while clicks rise and conversion rate drops, Google may be finding cheaper but less qualified traffic. Those problems require different fixes.
4. Separate budget limits from auction competitiveness
Lost impression share to budget and lost impression share to rank should not be treated as the same thing. A campaign losing demand because of budget may have room to scale. A campaign paying more while losing more to rank may need a closer review of Target CPA or Target ROAS, profitability, query mix, and local competition before more budget is added.
5. Make one meaningful change at a time
Avoid changing the target, budget, campaign structure, conversion actions, and creative at the same time unless the business requires it. The fastest way to learn nothing from a platform change is to change five other things with it.
The Longer-Term Answer Cannot Be to Keep Paying Google More
Google remains one of the most powerful channels in marketing because it captures people when they express intent. That is also why so many growth plans have become overly dependent on it. Over time, high-intent auctions tend to get more expensive as competition grows and more decision-making moves into automated systems.
Diversifying the marketing mix does not mean taking 20% of Search and dropping it into the newest shiny channel. It means deciding what job each channel is supposed to do, then using incrementality to determine whether the investment created an outcome that would not have happened anyway.
- Search can capture existing demand and high-intent action.
- Paid social and video can create and shape demand before the search happens.
- Programmatic and emerging channels can extend reach, influence consideration, and create new discovery paths.
- CRM and first-party activation can nurture, convert, and retain people while improving the signals used across paid media.
- Incrementality testing can separate real contribution from conversions a platform can easily claim.
That is where Amsive’s Audience Science® approach matters: connecting audience understanding, channel roles, testing, and measurement so the next dollar follows incremental value, not simply the platform with the easiest attribution story.
What We Are Watching Next
This is the first article in an ongoing series. We will keep the study group and pre-change benchmark consistent while the auction settles, then report what continues, what returns to normal, and what changes enough to require action.
- Budget-constrained versus other Search campaigns: Do higher CPCs and greater rank loss continue?
- Target CPA versus Target ROAS: Does the first-week difference hold as more conversion and value data matures?
- Brand, Generic, and Competitor Search: Where is Google paying more for existing demand versus finding new demand?
- Broad, Phrase, and Exact match: Do cost changes and conversion-quality changes continue to differ by match type?
- Portfolio bidding and conversion density: Do shared signals or higher-volume campaigns settle faster?
- Lead and customer quality: Are platform conversions producing the same bookings, pipeline, revenue, and customer value?
- PMax, Shopping, and Demand Gen: What happens when we isolate campaigns using target-based bidding and meaningful value signals?
We are also interested in what other advertisers are seeing. Are budget-limited campaigns moving closer to their targets? Are clicks getting cheaper while lead quality changes? Are some verticals settling faster than others?
The useful question is not whether the update was good or bad. It is where Google’s new bidding behavior is creating profitable scale, where it is simply creating more traffic, and whether the business signals being sent back to the platform are strong enough to tell the difference. We will keep watching and share what the next few weeks show.
How We Approached the Analysis
Each advertiser was compared with its own earlier performance using the same conversion definitions before and after the change. For lead-generation campaigns, we focused on CPA, conversion rate, volume, and lead quality. We used ROAS only where the conversion values represented meaningful business value and the campaigns were using value-based bidding.
We also excluded accounts that had major tracking, campaign structure, launch, or pause changes that would make the comparison unreliable. Before calling something a broader trend, we checked whether one unusually large account was skewing the result.
This is an observational read of what changed after August 17. It does not assume that the bidding update caused every change in performance.
Sources + Related Reading
Methodology Note: Amsive analysis of Google Ads data. Primary comparison: August 17–23, 2026 versus the weekly average from July 27–August 16, 2026. Match-type analysis compares August 17–23 with August 10–16. Results are observational, and sample sizes vary by segment. ROAS is used only where conversion values represent meaningful business value and value-based bidding is in place.
Stay Ahead of What’s Happening in Digital
Search is changing inside the auction and before people ever click. Knowing how Google interprets your targets matters, but so does understanding how AI, social, media, and third-party content now shape discovery.
Watch our latest webinar on demand, Earn Trust in AI Discovery with Search, Social, and Media, for practical ways to strengthen visibility as search behavior changes.
Want a closer look at your paid search performance? Connect with Amsive.
